Howdy,
i have a bit of a problem here. I have built a super regulator based on the articles published by Jung to go with the opamp preamp. I have attaached the schematic. Its a parts bin build so i had to be a little creative. I replaced the zener diode with 4 red LEDs and also added a JFET/BJT cascode. Opmp is the THS4631. At the moment its on the breadboard and is stable.
The pass element is a BD179 (pretty slow 3Mhz) but as i said, its a parts bin build.
Building this one up from simulation to breadboard was a fun exercise, boy i learned a lot about correct grounding. Phase margin looks OK with ESR > 0.1 Ohm for the output cap in simulation, in practice i have no good ESR meter but after correcting the grounding scheme its now spot on and stable. I hope it still is when i move it over to perfboard.
So no problems right? Well my problem is tht i cant really measure its performance. I have an old analog Leader 60Mhz scope that can go down to 500uV per division when i use the 10x magnifier option. Noise seems to be below that. So i calibrated my crappy onboard soundcard to be 0dbV with 1V rms input and fired up Baudline, a nice high quality spectrum analyzer. Connected the PS via a big 10uF DC blocking cap. Noise floor is at -110dBV and doesnt change with or without regullator connected.
So how does one measure the performance of such ultra low noise stuff? Especially on a budget. Searching through the internet i found several R&S psophometers for sale, the Rohde & Schwarz UPGR seems a good choice since it can measure RMS voltages down to 3uV.
My Multimeter is a fluke 87 IV, it can measure (theoretically) down to 1uv AC but connected to the PS it shows 28uV, and its specs show that at this levels its not to be trusted anymore, although i can see different values when i change the grounding path on the breadboard.
So, is it best to invest in a good quality Psophometer or should i just wire up a Dual Opamp (i dont really have any low noise ones left) to a 40db or 60db gain passband Filter?
To test transient response i wiill us a square wave generator / Opamp / Mosfet current sink and see if there is any impact on the load regulation and ringing.
And my final question is, my preamp design is opamp based, are supply coupling caps still needed when the regulator has an output Z in the microohm range? Or do we reach the point where a cap, especially large elkos do more harm then good, looking at the impedance curves of capacitors...
My preamp is also based on a Jung Design and can be found here: http://www.diyaudio.com/forums/analog-line-level/286260-baby-first-opamp-buffer-preamp.html
The schematic is not up to date, the input pot has been changed and accordingly some resistors values of the compensation network and feedback path have been lowered for less gain and more importantly less noise.
The regulator is designed for 100ma current draw max although the complete stereo preamp circuit will draw roughly 30ma max.
i have a bit of a problem here. I have built a super regulator based on the articles published by Jung to go with the opamp preamp. I have attaached the schematic. Its a parts bin build so i had to be a little creative. I replaced the zener diode with 4 red LEDs and also added a JFET/BJT cascode. Opmp is the THS4631. At the moment its on the breadboard and is stable.
The pass element is a BD179 (pretty slow 3Mhz) but as i said, its a parts bin build.
Building this one up from simulation to breadboard was a fun exercise, boy i learned a lot about correct grounding. Phase margin looks OK with ESR > 0.1 Ohm for the output cap in simulation, in practice i have no good ESR meter but after correcting the grounding scheme its now spot on and stable. I hope it still is when i move it over to perfboard.
So no problems right? Well my problem is tht i cant really measure its performance. I have an old analog Leader 60Mhz scope that can go down to 500uV per division when i use the 10x magnifier option. Noise seems to be below that. So i calibrated my crappy onboard soundcard to be 0dbV with 1V rms input and fired up Baudline, a nice high quality spectrum analyzer. Connected the PS via a big 10uF DC blocking cap. Noise floor is at -110dBV and doesnt change with or without regullator connected.
So how does one measure the performance of such ultra low noise stuff? Especially on a budget. Searching through the internet i found several R&S psophometers for sale, the Rohde & Schwarz UPGR seems a good choice since it can measure RMS voltages down to 3uV.
My Multimeter is a fluke 87 IV, it can measure (theoretically) down to 1uv AC but connected to the PS it shows 28uV, and its specs show that at this levels its not to be trusted anymore, although i can see different values when i change the grounding path on the breadboard.
So, is it best to invest in a good quality Psophometer or should i just wire up a Dual Opamp (i dont really have any low noise ones left) to a 40db or 60db gain passband Filter?
To test transient response i wiill us a square wave generator / Opamp / Mosfet current sink and see if there is any impact on the load regulation and ringing.
And my final question is, my preamp design is opamp based, are supply coupling caps still needed when the regulator has an output Z in the microohm range? Or do we reach the point where a cap, especially large elkos do more harm then good, looking at the impedance curves of capacitors...
My preamp is also based on a Jung Design and can be found here: http://www.diyaudio.com/forums/analog-line-level/286260-baby-first-opamp-buffer-preamp.html
The schematic is not up to date, the input pot has been changed and accordingly some resistors values of the compensation network and feedback path have been lowered for less gain and more importantly less noise.
The regulator is designed for 100ma current draw max although the complete stereo preamp circuit will draw roughly 30ma max.
Attachments
One option would be to build a 40 or 60dB low noise test amp, there are designs floating around here, this will be useful for years to come.
Also, connect your sound card to the psu output through a resistor to avoid shorting it. Use the lowest Rload the soundcard allows. You then know the current into the supply and if you can measure the voltage you know the Rout.
Jan
Last edited:
I know the feels, man. You will need a pulsed current sink to measure transient response. A high gain (1000X or so) bandwidth limiting (20hz-20K or whatever) amplifier to measure input rejection and noise. Its best if you have FFT capability on your scope so you can differentiate input rejection from noise. What you may find is that its hard to differentiate your input rejection signal from the rest of the output noise.
Here is a datasheet for a discrete regulator that has the test circuits for these things that show component values and resulting scope shots and plots. It may help you out with coming up with the test methods.
Discrete Voltage Regulator Full Data Sheet - Sparkos Labs.com
Here is a datasheet for a discrete regulator that has the test circuits for these things that show component values and resulting scope shots and plots. It may help you out with coming up with the test methods.
Discrete Voltage Regulator Full Data Sheet - Sparkos Labs.com
- Status
- Not open for further replies.